Any meaningful UX testing in 2023 needs to account for ad blockers and password managers.
If your site or app doesn’t work with popular ad blockers, or it refuses to allow logins pasted in from password managers, it’s broken.
Yes, some executive type will want to argue about this because they think ad blocking will go away or they misunderstood some now outdated infosec guidance. They’re wrong. Users use ad blockers and password managers and if your stuff doesn’t work with them, it’s broken.
+1 to all of that ... additionally
Your content is not sufficiently special that I can't get it from a lower-lift source.
Anyone know of a browser extension that reminds you why you don't want to go to that site instead of actually taking you there?
"I've got a little change in my pocket .... " as it were.
@tim_lavoie @tommorris One argument I've heard for using an external ad server is as an audit mechanism for the number of impressions/clicks.
If you do everything in-house, then the advertisers need to take your word for it that their ad was actually displayed.
@jamesh @tommorris Does an advertiser get this today from Google or Facebook? Is there any assurance that the ad was viewed by a person?
Even still, none of this requires surveillance-based ads, which are the dominant variety today. Opt out.
@tim_lavoie @tommorris For a website, their ad space is more valuable if there is more certainty about how many times they've been viewed. And for the advertiser, it's nice if theres fewer people to audit or trust.
This isn't saying that the current mess of third party cookie user tracking is necessary: just that there are reasons besides cross-site user tracking why a site might outsource ad serving.
@azonenberg @tommorris It's more complicated than that. Adblockers often block tracking scripts as well, or even just telemetry scripts that are relatively harmless (but *can* be used for tracking). These aren't actually business-critical on a per-user basis, but it's possible for the site's main scripts to break by accident if the telemetry library isn't there.
(Naturally, I use an adblocker on my work computer, which means if the site breaks for me, I'll know about it and will be able to give a heads-up to whoever wrote the overly-entangled code.)
@azonenberg Hmm, I may not have been clear enough.
Imagine a piece of JS that has business logic, maybe making a fetch request for more widgets to show on the screen. If it detects a certain condition, it calls NewRelic.recordEvent(...) and goes on about its business. This isn't business critical, and if you asked the developer "hey, how important is it that this works?" they would say "eh, no big deal if it doesn't send the info". But what they're not anticipating is that window.NewRelic might be *undefined*. They've forgotten to include an if(NewRelic){...} around it.
That's the kind of scenario I mean. It comes up now and then.
It's broken, but it's not an *intentional* reliance on telemetry.
@[email protected] Hmm..."telemetry" covers a *huge* range of things. That's part of the problem here. So you have: - A/B testing and "user journey" analysis (more on the marketing and sales side) - Feature usage statistics (important for knowing what you need to keep supporting) - Error reporting (critical for knowing if you've broken the site) ...and all of that sometimes gets lumped under "telemetry". I work on the backend, and I rely very heavily on telemetry to know what's working, what's being used vs. what can be deprecated, whether a dark-launch is successful, etc. I don't know for sure what people do on the frontend, but I imagine it's very similar. And as long as it's not tracking individual users, I don't see the problem.
@azonenberg Hmm..."telemetry" covers a *huge* range of things. That's part of the problem here. So you have:
- A/B testing and "user journey" analysis (more on the marketing and sales side)
- Feature usage statistics (important for knowing what you need to keep supporting)
- Error reporting (critical for knowing if you've broken the site)
...and all of that sometimes gets lumped under "telemetry".
I work on the backend, and I rely very heavily on telemetry to know what's working, what's being used vs. what can be deprecated, whether a dark-launch is successful, etc. I don't know for sure what people do on the frontend, but I imagine it's very similar. And as long as it's not tracking individual users, I don't see the problem.
@tommorris Once Upon A Time there was a thing called "zonealarm" which took it upon itself to remove the "ADVERTISE" button when it displayed my web site to its users (aka victims).
Hence endless conversations along the lines of:
"So how do I actually place an advertisement on your advertising web site?"
"Press the ADVERTISE button and fill in the form."
"What ADVERTISE button?"
I do *not* consider that it was my duty to code around this "zonealarm" malware, so I'm going to disagree with you.
@TimWardCam
Cool. But don't complain when you site breaks because your users are protecting themselves.
Also? "This one time a single user tool broke something it shouldn't have and now I'm to try to punish all users using similar tools forever" is a bad look.
@tommorris
@dymaxion @tommorris There are thousands of tools out there. I'm supposed to buy subscriptions to all of them, and the platforms they run on, and test with everything once a week? For a free public service I provide?
Meanwhile, back in the real world ...
@dymaxion @tommorris How much does an iPhone cost?
It's a free public service. If it works for almost everybody who tries to use it, great. If someone complains that it's not working for them, I investigate.
@TimWardCam @dymaxion My intended suggestion was “if you are testing on a device, test with a popular adblocker installed (as quite a lot of users will have one)” not “extend testing scope beyond practicality”. The target of my post was professionals working on commercial projects with resources to test with and an expectation or obligation to test across multiple devices.
uBlock Origin with the default settings is free and available for most popular desktop browsers.
@tommorris See also: scroll-jacking. Interposing clipboard copies. Redirection through interstitial load. Useless captchas.
When a machine solves your human verification better than humans do, you're "totally winning that war".
The Sixth Replicant
Tom Morris
Brian Chamberlain
Jon Ribbens
apophis
Eleanor LNR Blair
🥥 Tucker Carlson's Nuts 🥥
Scissors Cut Paper...
Tim Lavoie
James Henstridge
Eleanor Saitta
stefan
Jay Little
Andrew Zonenberg
varx/tech
Grinning Cat 
🚫👑
David S
Scenario
AMS
Jen
Tim Ward ⭐🇪🇺🔶 #FBPE
Rev. GothAlice
Alex Michael Berry 🐈🐩
BoredomFestival
aerique